MathGroup Archive 2011

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Which Test is Used to Verify Assumptions in TTest

  • To: mathgroup at smc.vnet.net
  • Subject: [mg120194] Re: Which Test is Used to Verify Assumptions in TTest
  • From: Gregory Lypny <gregory.lypny at videotron.ca>
  • Date: Tue, 12 Jul 2011 07:00:08 -0400 (EDT)
  • References: <201107091132.HAA13524@smc.vnet.net> <4E1873F6.7040605@wolfram.com>

Thank you, Andy.  Much obliged.

Gregory


On Sat, Jul 9, 2011, at 11:29 AM, Andy Ross wrote:

> On 7/9/2011 6:32 AM, Gregory Lypny wrote:
>> Hello everyone,
>> 
>> 	I've run some simple t-tests of means using TTest and was pleased to see that Mathematica pretests for the validity of underlying assumptions.  In a very skewed sample, I received a warning that the normality assumption had likely been violated.  The warning looked like this:
>> 
>> 	TTest::nortst: "At least one of the p-values in 0., resulting from a test for normality, is below 0.05. The tests in T require that the data is normally distributed.
>> 
>> 	Is it possible to find out precisely which pretest Mathematica performed in order to generate that message?
>> 
>> Regards,
>> 
>> Gregory
>> 
>> 
>> 
> 
> In order to find out what test was being used apply DistributionFitTest to your original data and ask for the "AutomaticTest" property. This will give the name of the test that was being used.
> 
> Something like...
> 
> DistributionFitTest[ data, Automatic, "AutomaticTest"]
> 
> or if you just want to perform the test...
> 
> DistributionFitTest[data,Automatic,"TestDataTable"]
> 
> This process will be documented in future versions of Mathematica but in the meantime I hope this helps.
> 
> Andy Ross
> Wolfram Research



  • Prev by Date: Re: sequence of functions
  • Next by Date: Re: How to write a "proper" math document
  • Previous by thread: Re: Which Test is Used to Verify Assumptions in TTest
  • Next by thread: Calculating General Tensor Product of Gradients